INDEX

Explanations

walking heel-toe

exploring difficult and sensitive subjects

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

PanelVisual

0.51

 वडिला

0.51

Ё

0.50

ಡ

0.48

 ರೋ

0.46

abut

0.45

nego

0.45

 Lanc

0.45

rega

0.45

ipat

0.43

POSITIVE LOGITS

 ambienti

0.53

ذ

0.50

ون

0.47

جميع

0.47

 ambientes

0.46

भी

0.45

也都

0.45

 defensa

0.44

泓

0.41

ци

0.41

Activations Density 0.001%